View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Elsevier - Publisher Connector FEBS Letters 582 (2008) 2441–2444 Prediction of replication time zones at single nucleotide resolution in the human genome Feng Gao, Chun-Ting Zhang* Department of Physics, Tianjin University, Tianjin 300072, China Received 6 March 2008; revised 3 June 2008; accepted 4 June 2008 Available online 12 June 2008 Edited by Takashi Gojobori zones) themselves are not the minimum replication units; Abstract The human genome is structured at multiple levels: it is organized into a series of replication time zones, and mean- rather, most of them are composed of several tandem clustered while it is composed of isochores. Accumulating evidence sug- replicons [21]. It was reported that the RT-zones with higher gests a match between these two genome features. Based on G + C content replicate early, whereas the RT-zones with low- newly developed software GC-Profile, we obtained a complete er G + C content replicate late [6,7]. Many experiments re- coverage of the human genome by 3198 isochores with bound- vealed that the replication timing switch sites are consistent aries at single nucleotide resolution. Interestingly, the experi- with the transitions of G + C content along the chromosomes mentally confirmed replication timing sites in the regions of investigated [6,7]. Very recent experiment shows that two adja- 1p36.1, 6p21.32, 17q11.2 and 22q12.1 nearly all coincide with cent replication timing sites constitute an RT-zone, which is the determined isochore boundaries. The precise boundaries of exactly an isochore [7]. In the current work, isochores in the the 3198 isochores are available via the website: http://tubic.t- human genome were determined such that their boundaries ju.edu.cn/isomap/. Ó 2008 Federation of European Biochemical Societies. Pub- coincide with replication timing sites that have been experi- lished by Elsevier B.V. All rights reserved. mentally confirmed. Therefore, other isochore boundaries, in addition to those coincide with confirmed replication timing Keywords: Human genome; Isochore; Replication timing; sites, are possible to have the same role, thereby boosting the Segmentation; G + C content number of potential replication timing sites for further exper- imental validation. In our previous papers [16,17], only 56 isochores with length longer than 3000 kb were studied, which cover about 21% of the whole human genome, whereas in the present study, 3198 1. Introduction isochores with length longer than 30 kb were identified, which cover nearly 100% (except gaps) of the whole human genome. Since the pioneer work of Bernardi and co-workers, it has been well established that the mammalian genomes are com- posed of large sequence segments of fairly homogeneous 2. Materials and methods G + C content that was revealed by the analytical ultracentri- fugation of bulk DNA in the mid 1970s [1]. The long DNA The whole human genome sequences were downloaded from segments (on average >300 kb) of fairly homogeneous G + C http://hgdownload.cse.ucsc.edu/downloads.html#human released in content lately were given the name ÔisochoresÕ [2]. Since then, March 2006. Two independent methods were used to identify bound- aries of isochores. The first method is called the cumulative GC-pro- the issues of isochores in mammalian and other eukaryotic file (z0 curve) method [15]. For a given genome or chromosome, genomes have attracted wide attention [3–17]. there is a unique cumulative GC-profile or z0 curve corresponding One of challenging problems in isochore studies is to deter- to it. The z0 curve or the cumulative GC-profile is used interchange- mine how many isochores there are in the human genome, and ably in this paper. Note that the essence of cumulative GC-profile is the solution to this problem largely depends on the definition to intuitively display the variations of the G + C content along a genome or chromosome. It is not the G + C content itself. Rather, of isochores. Bernardi and co-workers recently obtained a the derivative of z0 curve with respect to the base position n is neg- complete coverage of the human genome by about 3200 isoch- atively proportional to the G + C content at the given position, i.e., ores [10,11], and they also found the rules according to which G+C µ Àdz0/dn. Therefore, the average slope of the z0 curve within a the 3200 isochores of the human genome are assembled in high region reflects the average G + C content of the sequence within this region. If the z0 curve in a region is an approximately straight line, resolution chromosome bands (850 bands) [12]. Being different the G + C content keeps approximately constant within this region. from Bernardi and co-workers [10,12], here we present an A jump (drop) in the z0 curve indicates a decrease (increase) of alternative solution. A large body of evidence has shown that the G + C content. A turning point in the z0 curve indicates a switch the mammalian chromosome is organized into a number of site, at which the G + C content undergoes an abrupt change from a large replication time zones [18,19], ranging from 100 kb to relatively GC-poor (GC-rich) region to a relatively GC-rich (GC- poor) region. The point at which the derivate of the z0 curve is 2 Mb in length roughly [20]. The replication time zones (RT- not continuous is called a turning point or segmentation point. However, a more convenient method to find the coordinates is to use a newly developed segmentation algorithm [22], which is imple- *Corresponding author. Fax: +86 22 2740 2697. mented using a computer program, called GC-Profile [23]. The E-mail addresses: [email protected] (C.-T. Zhang), ren_zhang@ya- cumulative GC-profile, the distribution of G + C content, the hoo.com (C.-T. Zhang). isochores and their boundary coordinates for each of the human 0014-5793/$34.00 Ó 2008 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2008.06.008 2442 F. Gao, C.-T. Zhang / FEBS Letters 582 (2008) 2441–2444 chromosomes as well as the features of isochores are available online region, which is just the transition site at which the G + C con- by visiting http://tubic.tju.edu.cn/isomap/. Consequently, the total tent varies from 37% to 51% between the isochore 17-74 number of isochores is 3198 for the whole human genome (hg18). (26,457,053–26,735,738) and the isochore 17-75 (26,735,739– 26,954,701). The two isochores identified here, i.e., the isocho- re 17-74 and isochore 17-75, are exactly the two isochores stud- 3. Results and discussion ied by Schmegner et al. [25]. The fourth evidence we show here concerns the replication 3.1. Comparisons with experimental evidence timing of the MN1/PITPNB gene region on chromosome 22 In what follows, we will show evidence that the boundaries [7]. The analysis of the G + C content for this fragment of of some isochores obtained here are in accordance with the DNA sequence showed that there are four distinct sub-regions known replication timing sites confirmed by experiments. with different G + C content. The proximal sub-region has a Please refer to the related materials (such as Supplementary G + C content of 50.1%, and the G + C content of the second, Figures 1–4) from http://tubic.tju.edu.cn/isomap/supplemen- third and fourth sub-regions are 39.5%, 53.0% and 39.5%, tary while reading the following text. The first evidence we respectively. The four sub-regions are termed A-, B- C- and show here concerns the replication timing switch site in the hu- D-isochores, respectively. Perhaps, the most surprising finding man major histocompatibility complex (MHC) sequence. One of the work is that the four isochores match the replication of the replication timing sites was found experimentally within timing zones with such a degree that the authors called the this sequence [6]. Readers are suggested to refer to the UCSC match ‘‘perfect’’ [7]. The authors found that the A- and C- Genome Browser in the region of chr6: 31,647,700–33,232,435, isochores replicate early, whereas the B- and D-isochores rep- or Supplementary Figure 1. The replication timing switch re- licate late during the S phase of the cell cycle. There are three gion is between the gene NOTCH4 (chr6: 32,270,599– sharp boundaries (transitions) of the four isochores, i.e., the 32,299,822) in MHC class III (isochore H3) and the gene boundaries between isochores A/B, B/C and C/D, respectively. HLA-DRA (chr6: 32,515,625–32,520,799) in MHC class II The analysis of the G + C content (using the window-based (isochore L2). To be more accurate, the switch region should method) showed that the transition occurs within a region of be located within the interval 32,280–32,320 kb (please refer few kb. Interestingly, the four replication zones or isochores to Fig. 6 or Fig. 8 in Ref. [6]), while the segmentation point ob- the authors studied [7] are basically the four isochores ob- tained by our software GC-Profile, 32,300,166 bp within the tained in this work. Please refer to the figure of UCSC Genome 6p21.32 band, is situated exactly within the above interval. Browser (hg18) in the region of chr22: 26,129,874–26,866,656, Therefore, the boundary between the predicted isochore 6-30 or Supplementary Figure 4. This region consists of four dis- and isochore 6-31 is precisely consistent with the replication tinct isochores, i.e., the isochores 22-32, 22-33, 22-34 and 22- timing site confirmed experimentally [6].